Big Tech's hidden debt of $1.65 trillion: The AI race on borrowed funds and melting cash flow
The five largest technology corporations — Alphabet, Microsoft, Amazon, Meta, and Oracle — have accumulated $1.65 trillion in hidden liabilities that are not reflected in their official balance sheets. These involve contracts for GPU leases, data centers, and joint venture obligations. Under current accounting rules, these amounts are not considered debt until the facilities are operational, creating an illusion of financial stability.
The Scale of Shadow Financing
The most striking example is Meta, whose hidden liabilities have reached $420 billion, three times its official debt. Oracle's situation is even more dramatic: its off-balance-sheet liabilities have grown 30-fold over the past four years. The Bank for International Settlements (BIS) has already characterized this practice as "shadow lending."
The problem is compounded by the fact that four of the five companies are reporting their financial results within the next two weeks. Official debt will appear moderate, while $1.65 trillion will remain in footnotes, never making it into report headlines.
Cash Flow Under Pressure from CapEx
Bank of America warns of an impending liquidity crisis. The free cash flow of the largest data center operators will drop from $191 billion in 2025 to $19 billion in 2026, and by 2027 it will turn negative at approximately -$26 billion. The reason is record capital expenditures on AI infrastructure, which have already exceeded $725 billion, while sector debt issuance will reach around $175 billion — six times the average over the past five years.
The long-term forecast, however, paints a different picture. By 2030, the combined free cash flow of Meta, Google, Microsoft, and Amazon could grow nearly tenfold to $650 billion. But this optimistic scenario is only possible if current investments pay off.
Betting on Monopoly
Corporations are betting not just on infrastructure returns, but on future monopoly. They aim to become an indispensable link for anyone using AI and to profit from every query. Alphabet appears to be the frontrunner: Google owns the entire AI chain, Google Cloud revenue grew 63% year-over-year, its order backlog nearly doubled to $462 billion, and it has started selling its proprietary TPU chips to external clients.
The worst-case scenario would occur with weak AI demand. Facilities would have to be written down, and losses would fall on investors and insurers who financed construction through private credit and project bonds, unaware of the full extent of the risks.
Expert commentary: The situation resembles the 2008 mortgage crisis, when complex financial instruments masked real risks. The difference is that now the bet is on technological monopoly rather than real estate. Investors should closely monitor reports and look not only at official balance sheets but also at footnotes — that's where trillions of dollars are hidden, which could crash the market if the trend reverses.